Insurance, Risk and Audit Advisory Case Studies
Professional5 Case StudiesUpdated Sep 2026
Insurance audit and advisory work tends to fail in unglamorous places: a reconciliation signed by the wrong person, a forecast resting on an assumption nobody retested, a reserve input table that outlived the product it described. The five cases below come from fictional carriers, agencies, and lenders, and each follows one decision from the first signal to a documented outcome. They are written for practitioners who already know the mechanics. The useful parts are the thresholds that triggered action, the evidence that was accepted or rejected, and the owners who had to live with the result.
Case Study 1
The ceded premium gap that would not close
Harbor Ridge Mutual, a regional property and casualty carrier with about 40,000 personal auto policies in force
The situation
At Harbor Ridge Mutual, the month-end close stalled when the ceded premium in the general ledger stopped matching the reinsurance cession statement. The ledger showed 398,120 in ceded written premium. The reinsurer's statement showed 412,660. Dana, the senior accountant in finance operations, had three source files open at once: the broker bordereaux, the cash receipts register, and a manual journal posted in the last two days of the period. Nothing in the close checklist explained the gap, and the reinsurer had already issued its statement with a payment due in the second week of the following month.
The challenge
The team needed to answer three questions before anyone could sign the close. Was the difference timing, error, or something that required a closer look at who posted the manual journal? Did the variance change the reinsurance recoverable or the unearned premium reserve? And at what size did the issue stop being a local correction and become a matter for the controller or the audit committee chair? The policy set those thresholds at 5,000 for controller escalation and 25,000 for committee notice, so the answer mattered.
Analysis
The first pass in this case compared the bordereaux line by line against the ledger, not against the explanation offered for the gap. That comparison isolated fourteen policy endorsements that had been ceded twice, once in the original batch and again through the manual journal. The duplicate pattern pointed toward a process weakness rather than a single bad entry. Each endorsement carried a different preparer initial, so the team could test whether the duplication clustered around one step in the workflow or spread across the group. It clustered around one step, which narrowed the search considerably.
The exposure was not the 14,540 alone. Ceded premium flows into the recoverable and into the reserve calculations, so a duplicated cession inflates the amount the company believes it can collect. The team weighed that against the control evidence. The close checklist had a sign-off box for the reconciliation, but the box had been ticked by the same person who prepared the journal. That is a segregation failure regardless of the amount. Against the thresholds, the confirmed portion sat below the controller line, but the segregation issue alone justified notice to the committee chair so the matter was on record before audit fieldwork began.
- 1Pull source recordsObtain bordereaux, ledger detail, and receipts for the same period without relying on summaries.
- 2Match line by lineIsolate each variance to an endorsement, batch, or manual entry.
- 3Test the causeDecide whether the difference is timing, duplication, or an error needing a correcting entry.
- 4Apply thresholdsCompare confirmed amounts with controller and committee limits before deciding who is told.
- 5Evidence the fixRecord the correction, the reviewer, and a retest date in the close file.
Actions taken
- Reverse the duplicated cession entries through a documented correcting journal, with the bordereaux line references attached so the reversal can be traced without a second investigation.
- Require a preparer and a separate reviewer for every manual journal touching ceded premium, and configure the ledger so the reviewer field cannot match the preparer identifier.
- Add an automated matching step that compares the bordereaux total to the ledger each business day during the close, flagging any variance above 5,000 to the controller within the same day.
- Send the committee chair a short written memo summarizing the cause, the corrected amount, and the control changes, then schedule a retest of the reconciliation control after two full closes.
Results
The correcting entries cleared the difference before the reinsurer's payment date, and the recoverable was restated by 14,540 in the close package. The segregation finding was logged as a control deficiency rather than a material weakness, mainly because the corrected amount was small and the duplicate pattern was confined to one workflow. The retest after two closes showed no repeat duplication. The daily matching step also caught one unrelated timing difference of 2,310 that cleared the following morning. The team's main regret was that the manual journal path had existed for years without a reviewer requirement.
Questions for discussion
- Which of the three thresholds in this case would you have revisited first, and what evidence would have changed your view of it?
- If the duplicate pattern had clustered around a single preparer rather than a workflow step, how would the remediation and the committee communication change?
- How would you decide whether a control that operated correctly for several closes still deserved a retest after a process change?
Case Study 2
Cash flow shortfall in a growing agency book
Lakeshore Specialty Risk, a managing general agency placing commercial liability and inland marine business for about 180 independent agencies
The situation
Lakeshore Specialty Risk had planned for commission income to rise by 12 percent over the year, funding a larger staff and a new underwriting desk. By the end of the second quarter, commission income ran only about 4 percent ahead of the prior year, and operating cash was 1,240,000 below the plan line. Ferhat, the finance lead, had built the forecast on an 88 percent renewal retention rate drawn from three prior renewal cycles. Retention had slipped to 81 percent in the most recent cycle, and two large agencies had moved a combined block of accounts to a competing wholesaler. The board wanted a revised forecast and a recommendation before the next hiring round was approved.
The challenge
The question was not simply how much the forecast had missed. The team needed scenarios that showed when cash would reach the minimum liquidity covenant, which the lender set at 900,000, and which levers were available without breaking agency contracts. Some assumptions were stale, some were genuinely uncertain, and the board had asked for a response it could defend if the next two renewal cycles went either way.
Analysis
Teams grounded in Insurance, Risk and Audit Advisory usually separate a forecast into the parts that were observed and the parts that were assumed, then test each assumption against the records that would contradict it. For this agency, the observed items were booked commissions, paid remittances, and the timing of carrier statements. The assumptions were retention, new business growth, and the average commission rate on new placements. Retention was the most sensitive, because a one point change moved the annual commission base by a meaningful amount. The monthly bordereaux made it possible to check the retention trend against individual agency books rather than a blended average.
The team built three scenarios. The base case assumed retention recovered to 85 percent. The downside held it at 81 percent with the lost block gone. A severe case added a sixty-day delay in carrier commission payments. Each scenario was tested against the 900,000 covenant floor and against a second threshold the board had set, which was a cash buffer of 300,000 above the floor before any new hiring could be committed. Under the downside case, the buffer disappeared in the fourth month of the next fiscal year, which drove the recommendation. Control evidence mattered here too. The carrier statements were the only source that could confirm payment timing, so the team reconciled them against the remittance log before accepting the severe case.
- Early in the quarterRetention drop surfaces in the monthly bordereaux for two large agencies.
- Quarter closeCommission income and operating cash both fall below plan lines.
- Within the same weekTeam rebuilds three scenarios from the remittance log and carrier statements.
- Before the hiring roundBoard receives the scenarios and approves the pause with monitoring triggers.
- Following renewal cycleRetention and returned accounts are checked against the downside threshold.
Actions taken
- Freeze the new underwriting desk hires until retention holds at or above 84 percent for two consecutive monthly bordereaux, and tie each later hire to a named book of business.
- Open a retention review with the two agencies that moved accounts, focusing on service turnaround and pricing, and record each conversation with an outcome and a follow-up date in the account file.
- Agree a monthly cash checkpoint with the lender that reports the buffer above the 900,000 floor, and trigger a pause on discretionary spend if the buffer falls below 300,000.
- Rebuild the forecast each month from the bordereaux rather than the annual plan, and keep the three scenarios in one workbook so that changes to any assumption are visible to the board.
Results
The board approved a six-month hiring pause for the desk and kept the existing staff. One of the two agencies returned about 40 percent of the moved block within the following renewal cycle, after service changes were agreed. Retention finished the year at 84 percent, which sat between the base and downside cases, and the buffer never fell below 300,000. The lender accepted the monthly checkpoint without altering the covenant. The forecast had been wrong about the level of retention, but the scenarios had made the cost of being wrong visible early enough to act on.
Questions for discussion
- Which assumption in this forecast would you have tested first, and what record would have shown you it was drifting before the cash gap appeared?
- How would the recommendation change if the lender had refused the monthly checkpoint and set a tighter floor?
- When a retention decline is concentrated in a few large accounts, how much weight should a blended average still carry in the forecast?
Case Study 3
Third party premium payments that did not add up
Corvane Indemnity Group, a surplus lines carrier writing excess liability for hospitality and transportation accounts
The situation
Corvane Indemnity Group received a single premium payment of 486,000 for a new commercial umbrella policy. The payment came from a bank account in the name of a holding company that was not the named insured, and the broker's submission listed a different remitting entity. Within the same week, the policy was flagged for cancellation, with a return of premium requested to a third account. Ines, the compliance analyst, pulled the transaction into the monthly review queue after the premium finance team noted the mismatch. The indicators were present, but none was conclusive on its own.
The challenge
The team had to decide whether the payment pattern required a suspicious activity review, whether the documentation was adequate to explain the source of funds, and whether the broker relationship needed a restriction. The carrier's procedures set a review trigger for any premium payment from an unrelated third party above 25,000, and required escalation to the compliance officer within two business days of identification. The clock had already started.
Analysis
A single odd payment is a question, and a pattern of odd payments is a finding, so the first task was to build the full record of what changed and when. The record showed the premium payment, the mismatched remitter, the cancellation request, and the return destination, all within a short window. The exposure was not the premium itself, since the carrier had already received it. The real risk was that the company was being used to move funds through a policy that could be unwound. The documentation was also weak. The broker had supplied a source of funds statement, but it was unsigned and did not name the holding company's beneficial owner.
The procedures distinguished between a documentation gap and a suspicion. A documentation gap could be cured by obtaining the signed statement, the beneficial ownership certificate, and a written explanation from the broker. A suspicion required a documented assessment, a decision by the compliance officer, and, where the rules applied, a report to the relevant authority within the deadline. The control evidence was thin on one point. The system had accepted the premium without a remitter match check, which was the control that should have stopped it at intake. That gap mattered regardless of how the case resolved.
Actions taken
- Hold the return of premium and the cancellation request until compliance has reviewed the remitter, and notify the broker in writing that the payment cannot be refunded to a third account.
- Obtain a signed source of funds statement and a beneficial ownership certificate for the holding company, and verify both against the payer's public registration before accepting the explanation.
- Prepare a written suspicion assessment that lists each indicator, the evidence for and against it, and the decision reached, then route it to the compliance officer for a documented determination.
- Add a remitter match check at premium intake that blocks acceptance when the payer differs from the named insured or the broker submission, with an override that only the compliance officer can approve.
Results
The broker produced the signed statement after the second request, but the beneficial owner certificate named an individual who could not be verified in the registry, so the compliance officer treated the matter as a suspicion. A report was filed within the required window, and the policy was restricted pending the outcome. Corvane did not refund the premium to the third account. The intake control went live within the month, and in the following quarter it blocked two similar payments, both of which were corrected by the insured's own bank. The file now shows how the decision was reached, which mattered more to the examiner than the outcome.
Questions for discussion
- At what point would you have stopped treating the documentation gap as curable and moved the matter into a suspicion assessment?
- How should a carrier weigh a broker's long relationship against an unverified payer when the premium is already in hand?
- What would a remitter match control need to log so that an override can be reviewed later without reconstructing the decision?
Case Study 4
Acquisition loan for a brokerage with thin margins
Tamarack Credit Partners, a private lender with a middle market fund focused on insurance distribution businesses
The situation
Redwood Coast Insurance Services, a brokerage with offices in four states and about 2,100 active commercial clients, asked Tamarack for a 6,500,000 term loan to acquire a smaller agency. The seller's trailing financials showed revenue of 9,800,000 and an EBITDA margin of 17 percent, but those figures were management prepared and had not been reviewed. Marisol, the credit analyst, received the request along with the seller's book of business summary, the buyer's three-year statements, and a schedule of contingent commissions. The target's largest carrier relationship accounted for roughly 38 percent of commission income, and that relationship was up for renewal in the coming year.
The challenge
The committee needed to know whether the combined business could service the debt under a downside case, whether the earnings were durable enough to justify the leverage, and what conditions would make the loan acceptable. Tamarack's policy capped senior leverage at 3.25 times pro forma EBITDA and required a debt service coverage ratio of at least 1.35 in the downside scenario. The request sat just above both limits in the base case.
Analysis
The credit file was rebuilt from the underlying records before any adjusted figure was accepted, because agency earnings depend heavily on contingent commissions and carrier relationships that a summary can easily hide. Marisol rebuilt the target's commission income from the carrier statements and found that contingent payments had contributed about 11 percent of the seller's reported EBITDA. Those payments depend on loss ratios and volume targets set by carriers, so they are not contractual. The seller's add-backs, including an owner compensation adjustment and a one-time technology cost, were partly supported by invoices and partly by estimates. The reviewed EBITDA came in about 9 percent below the seller's figure.
Three downside cases were run. The first removed the largest carrier relationship at renewal, with a 30 percent loss of its commissions. The second reduced new business volume by a fifth for two years. The third combined both with a 90-day delay in collections. Under the first case, debt service coverage fell to 1.12, below the policy floor, while the second stayed at 1.31. Suitability mattered as well. The borrower was a sponsor with a track record of two prior acquisitions, and the fund's mandate allowed a loan of this size only with a covenant package that gave Tamarack a cure right on carrier concentration.
- Verify earnings Rebuild commissions and add-backs from carrier statements and invoices, not the seller summary.
- Stress the downside Run carrier loss, volume, and collection delay cases against the 1.35 coverage floor.
- Confirm suitability Check sponsor record, mandate fit, and whether the covenant package addresses concentration.
- Set conditions Approve only with a concentration cure right and leverage step-down tied to renewal.
Actions taken
- Rebuild trailing EBITDA from the carrier statements and invoices for the last three years, and present the seller's adjusted figure only as a reconciling item below the reviewed number.
- Require a renewal status letter from the largest carrier and a schedule of contingent commission terms before final credit approval, and size the loan against the downside coverage case.
- Set senior leverage at no more than 2.75 times reviewed EBITDA at closing, stepping down to 2.5 times if the top carrier relationship is not renewed within the agreed window.
- Add a cure right tied to carrier concentration, an excess cash sweep that applies if coverage falls below 1.25, and quarterly delivery of carrier statements to the lender.
Results
The committee approved a reduced facility of 5,200,000 with a 2.75 times leverage cap at closing and a step-down condition tied to the carrier renewal. The seller accepted a lower price after the reviewed earnings were discussed, and part of the consideration was deferred against contingent commissions. In the first year, the largest carrier renewed at a lower share, and the sweep was triggered once, which the borrower met without a covenant breach. The loan performed as underwritten, but the team recorded that the original request had leaned on the seller's adjusted earnings more than the records supported.
Questions for discussion
- Which of the three downside cases would you have weighted most heavily, and would your answer change if the carrier had renewed early?
- How much of the seller's add-backs should a lender accept when the invoices support only part of them?
- When the sponsor's track record is strong but the target's earnings are concentrated, which condition would you insist on first?
Case Study 5
Five audit findings ranked for closure
Sable Creek Life Assurance, a life and annuity company with about 310,000 in-force contracts across three product lines
The situation
Sable Creek Life Assurance completed an internal audit of its annuity operations, covering premium processing, surrender payments, and reserve inputs. The review sampled 180 surrender transactions and 60 reserve input files from the period. Of the sampled surrenders, 14 had payment amounts that did not match the contract's surrender value schedule, and 9 of those were paid late against the policy's stated timing. Reserve input files showed two actuarial assumption tables that had not been updated after a product change. Sipho, the audit manager, had to decide how to report the results to the audit committee and the business owners.
The challenge
Management had five open items, several of them overlapping, and no common scale for ranking them. The committee wanted to know which findings could affect reported reserves or customer payments, who owned each fix, and what evidence would show the fix had worked. The threshold for a material misstatement was set by the finance team at a level tied to statutory capital, and two of the findings sat close to it.
Analysis
The ranking rested on what each finding could do to reported figures and customer outcomes, not on how many exceptions it produced. A high count of minor errors can hide a single item that moves a reserve. Sipho scored each finding on the size of the affected balance, the likelihood of recurrence, and whether the error ran in one direction. The surrender pricing error was systematic, with underpayments in 14 of 180 sampled cases, and the sampled amounts projected to a value near the lower end of the materiality band. The stale assumption tables were less frequent in the sample but touched every reserve calculation for the affected product, so they ranked above the late payments despite having fewer exceptions.
The closure tests were defined before owners started work. For the pricing error, closure required a corrected schedule, a re-performance on a fresh sample of at least 60 surrenders with zero exceptions, and a look-back to identify any customers underpaid in prior periods. For the assumption tables, closure required an actuarial sign-off with the revised tables, a reconciliation of the reserve output to the prior run, and evidence that change control now records the table version. Any finding projected above the materiality line went to the audit committee immediately. The committee received a status note only when a closure test had been performed, not when a fix had merely been reported.
Actions taken
- Assign the pricing finding to the head of annuity operations, with the chief actuary as co-owner, and set the corrected schedule as the first closure checkpoint.
- Run a look-back over all surrender payments in the prior two years to identify underpaid customers, and pay any shortfall with interest under the policy terms before the re-performance sample is drawn.
- Require actuarial sign-off and a reconciliation to the prior reserve run before any updated assumption table is released, and record the table version in the reserve output.
- Close the documentation findings only when the next review cycle samples the affected files and finds complete sign-offs, and report each closure test result in the status note to the committee.
Results
The surrender pricing error was corrected, and the look-back identified 22 underpaid customers whose shortfalls were paid within the period. The re-performance sample of 60 surrenders returned no exceptions, and the pricing closure was accepted by the committee in the following cycle. The stale tables were updated with actuarial sign-off, and the reserve reconciliation showed a movement that was explained and documented. Two of the five findings closed on the first retest. The late payment issue took longer, because the scheduling change depended on a system release, and it remained open at the next committee review.
Questions for discussion
- How would you weigh a finding that is rare in the sample but touches every balance against one that recurs often but affects small amounts?
- What evidence would you require before accepting that an underpaid customer population had been fully identified in a look-back?
- When a remediation depends on a system release outside the owner's control, how should the closure test and the committee reporting be adjusted?
Learning Resources
Professional CASE STUDIES
China, India, United States, Indonesia, Pakistan, Nigeria, Brazil, Bangladesh, Russia, Japan, Mexico, Philippines, Vietnam, Egypt, Turkey, Germany, United Kingdom, France, Italy, Canada, South Korea, Australia, Spain.